A Quest for Optimization of Postoperative Triage After Major Surgery

Innovative strategies to reduce costs while maintaining patient satisfaction and improving delivery of care are greatly needed in the setting of rapidly rising health care expenditure. Intensive care units (ICUs) represent a significant proportion of health care costs due to their high resources utilization. Currently, the decision to admit a patient to the ICU lacks standardization because of the lack of evidence-based admission criteria. The objective of our research is to develop a prediction model that can help the physician in the clinical decision-making of postoperative triage.

Our group identified a list of index events that commonly grants admission to the ICU independently of the hospital system. We analyzed correlation among 200 quantitative and semiquantitative variables for each patient in the study using a decision tree modeling (DTM). In addition, we validated the DTM against explanatory models, such as bivariate analysis, multiple logistic regression, and least absolute shrinkage and selection operator.

Unlike explanatory modeling, DTM has several unique strengths: tree models are easy to interpret, the analysis can examine hundreds of variables at once, and offer insight into variable relative importance. In a retrospective analysis, we found that DTM was more accurate at predicting need for intensive care compared with current clinical practice.

DTM and predictive modeling may enhance postoperative triage decision-making. Future areas of research include larger retrospective analyses and prospective observational studies that can lead to an improved clinical practice and better resources utilization.